Rational redesign of the active site of selenosubtilisin with strongly enhanced glutathione peroxidase activity

Journal of Catalysis 359 (2018) 27–35

Contents lists available at ScienceDirect

Journal of Catalysis journal homepage: www.elsevier.com/locate/jcat

Rational redesign of the active site of selenosubtilisin with strongly enhanced glutathione peroxidase activity Tingting Wang a, Jing Li b, Jiayun Xu a, Xiaotong Fan a, Linlu Zhao a, Shanpeng Qiao a, Tiezheng Pan a, Junqiu Liu a,⇑ a b

State Key Laboratory of Supramolecular Structure and Materials, College of Chemistry, Jilin University, 2699 Qianjin Road, Changchun 130012, China Beijing Bo Kang Kin Gene Technology Co., Ltd, Changping District Spark Street, Beijing 102299, China

a r t i c l e

i n f o

Article history: Received 1 July 2017 Revised 23 November 2017 Accepted 8 December 2017

Keywords: Enzyme redesign Selenoenzyme Peroxidase activities Selenium Artificial enzyme

a b s t r a c t The search for a perfect model to mimic the properties of the selenoenzyme glutathione peroxidase (GPx) has inspired great interest. Rational design and redesign of the structure–function relationship has become an indispensable technique. In this report, the active site of selenosubtilisin was successfully rebuilt by transferring the catalytically essential residue selenocysteine (Sec) to the edge of the substrate-binding pocket of the enzyme by artificial manipulation. Founding on computer-aided molecular simulation, the amino acid residue at position 63 (Ser in the wild-type enzyme) was selectively replaced with Sec using a cysteine auxotrophic expression system. The novel seleno63-subtilisin E gave a prominent 100-fold higher efficiency than the original seleno221-subtilisin E for GPx activity. Moreover, this seleno63-subtilisin E also had efficient hydrolase activity. Ó 2017 Elsevier Inc. All rights reserved.

1. Introduction Enzyme redesign to optimize the catalytic properties and understanding of directed evolution are a promising and challenging investigation [1]. It includes changes in substrate specificity, introducing new catalytic functions into the same active site, and conversion of ligand-binding sites into catalytic centers [2,3]. With the advent of tools for enzyme engineering, such as site-directed mutagenesis and three-dimensional structure prediction by computer-aided molecular simulation, active site redesign is gaining importance as a promising starting point for the rational redesign of enzymes [4]. Glutathione peroxidase (GPx, EC1. 11.1.9) is a structurally and functionally well studied selenoenzyme that catalyzes the reduction of hydroperoxides (ROOH) by glutathione (GSH) [5,6]. It is thought to be an important antioxidant enzyme in the prevention of lipid peroxidation and the corresponding disruption of membrane function. Enormous effort has been made in the development of artificial GPx models to explore its structure–function relationships [7,8], for example, introducing a catalytic center into an existing or artificially generated substrate-binding scaffold by chemical or genetic strategies [9]. However, single chemical modification of the enzyme’s primary catalytic group or alteration of ⇑ Corresponding author. E-mail address: [email protected] (J. Liu). https://doi.org/10.1016/j.jcat.2017.12.006 0021-9517/Ó 2017 Elsevier Inc. All rights reserved.

some vicinal amino acid residues of the essential active site usually having some shortcomings resulting in not fulfilling perfectly efficient enzyme models. Selenosubtilisin is the first GPx-like selenoenzyme obtained by chemical conversion of catalytic Ser221 to selenocysteine (Sec) to mimic GPx [10,11]. Our group has also generated it successfully using a cysteine auxotrophic expression system and has achieved high yield and catalytic efficiency [12]. The selenium side chain of Sec221 in selenosubtilisin is buried in a much deeper pocket and is presumably less accessible than its natural counterpart to hydroperoxides. Thus, this selenium-containing enzyme could not adopt the natural counterpart GSH of GPx but employs the aromatic donor 3-carboxy-4-nitroben-zenethethiol (ArSH) as ta reduced substrate. The GPx activity for enzyme-catalyzed reduction of H2O2 by ArSH is 4 lmol min
1 lmol
1, which is much lower than the native GPx of 5780 lmol min
1 lmol
1 [13]. Based on the principle of rational enzyme redesign, combined with computer-aided molecular simulation, we provide insight into biological function and mechanism in this investigation via free artificial manipulation of the essential catalytic group in the active site of enzymes to enhance the enzyme activity [14]. We redesigned active-site-directed mutation by first transferring the catalytically essential Sec to the edge of the substrate-binding pocket of the enzyme [15]. Depending on automated molecular docking by computer and the principle of energy minimization, the Ser63 was selectively replaced with Sec to be a novel GPx

28

T. Wang et al. / Journal of Catalysis 359 (2018) 27–35

mimic seleno63-subtilisin E [16]. Indeed, this alteration would allow the other substrate molecule to touch the catalytic site more easily and consequently enhance the catalytic efficiency. To this end, the novel seleno63-subtilisin E was generated using a cysteine auxotrophic expression system. Its GPx activity was substantially increased in comparison to that of seleno221-subtilisin E, and it retained efficient hydrolase activity. 2. Materials and methods

551 ACCAAAGAGC TTCATTCTCC AGCGCAGGTT CTGAGCTTGA CATGATGGCT 601 CCTGGCGTGT CCATCCAAAG CACACTTCCT GGAGGCACAA ACGGCGCTTA 651 TAACGGAACG TGCATGGCGA CTCCTCACGT TGCCGGAGCA GCAGCGTTAA 701 TTCTTTCTAA GCACCCGACT TGGACAAACG CGCAAGTCCG TGATCGTTTA 801 AATCAACGTA CAAGCAGCTG CACAA

2.1. Materials

2.2. Rational enzyme redesign by computational analysis

Plasmids and strains used in this study are listed in Table 1. Plasmid pSUB1 was kindly provided by Masayori Inouye of the University of Medicine and Dentistry of New Jersey, and the strains Escherichia coli DH5a and E. coli BL21DE3 were kept by our laboratory. The cysteine auxotrophic strain E. coli BL21cysE51 was a generous gift from August Bock of the Lehrstuhl fur Mikrobiologie der Universität München. Site-directed mutagenesis was carried out directly on the plasmid pET11a-prosubtilisin E, and the mutations were confirmed by DNA sequencing. A synthetic oligonucleotide

All studies were performed on an SGI O3800 workstation. They consisted of a binding-site module, flexible docking, and a threestep Ludi scores. In this study, identification of enzyme active sites and binding sites used ActiveSite-Search by locating cavities in the subtilisin E structure. When the search was completed, the largest site was automatically displayed on the structure using AsiteDisplay. The results were used to guide the protein–ligand docking experiments. By means of the 3D structures of subtilisin E and ArSH, which are built and optimized through the InsightII/Builder program, automated molecular docking was performed using the docking program Affinity. The potential function of the complexes was assigned using a consistent-valence force field (CVFF) and nonbonding interaction was dealt with by the cell multipole approach. To account for the solvent effect, the centered enzyme–ligand complexes were solvated in a sphere of TIP3P water molecules with radius 10 Å. Finally, the docked complex of the receptor with the ligand was selected by the criteria of interacting energy and geometrical matching quality. The Ludi method screens a large number of compounds and analyzes the geometrical fit of given chemicals into the binding site. It also can determine other good binding properties such as hydrogen bonds, lipophilic interactions, ionic interactions, and acyclic interactions. In general, a higher Ludi score represents a higher affinity and stronger binding of a ligand to the receptor.

(50 -CTTATAACGGAACGTGCA TGGCGACTCCTCA-30 ) has been employed as two primers to introduce an active-site (Ser221 ? Cys221) mutation into the wild-type subtilisin E gene. The other synthetic oligonucleotides, 50 -ACCAGGACGGCAGTTGTCACGGTAC GCATGT-30 , were for mutating Ser63 to Cys63. All chemical reagents and solvents were obtained from Sigma-Aldrich. The DNA sequence of subtilisinE(S221C) is as follows: 1 GCGCAAAGCT TTCCTTATGG CATTTCTCAA ATTAAAGCGC CGGCTCTTCA 51 CTCTCAAGGC TACACAGGCT CTAACGTAAA AGTAGCTGTT ATCGACAGCG 101 GAATTGACTC TTCTCATCCT GACTTAAACG TCAGAGGCGG AGCTAGCTTC 151 GTACCTTCTG AAACAAACCC ATACCAGGAC GGCAGTTCTC ACGGTACGCA 201 TGTAGCCGGT ACGATTGCCG CTCTTAATAA CTCAATCGGT GTTCTGGGCG 251 TAGCGCCAAG CGCATCATTA TATGCAGTAA AAGTGCTTGA TTCAACAGGA 301 AGCGGCCAAT ATAGCTGGAT TATTAACGGC ATTGAGTGGG CCATTTCCAA 351 CAATATGGAT GTTATCAACA TGAGCCTTGG CGGACCTACT GGTTCTACAG 401 CGCTGAAAA AGTCGTTGAC AAAGCCGTTT CCAGCGGTAT CGTCGTTGCT 451 GCCGCAGCCG GAAACGAAGG TTCATCCGGA AGCACAAGCA CAGTCGGCTA 501 CCCTGCAAAA TATCCTTCTA CTATTGCAGT AGGTGCGGTA AACAGCAGCA

Table 1 Plasmids and strains. Plasmids pET11a pSUB2 Strains TG1 DH5a BL21(DE3) BL21cysE51

T7 expression vector, Apr pET11a carrying wild-type subtilisin E SupE hsdD5 thia D(lac-pro AB) F0 [tra D36 pro AB + lac I q lac ZDM 5] SupE44 DlaC U169 (u80 lac ZDM 15) hsdR17 recA1 endA1 gyrA96 thi-1 rel1 hsdS gal (lcIts857 indI sam7 nin5 lacUV5-T7 gene 1) BL21(DE3) selB: kan cysE51

2.3. Expression and purification of two selenosubtilisin E Plasmid I (pET11a carrying mutant gene S221C) and plasmid II (pET11a carrying mutant gene S63C) were respectively transformed into strain BL21cysE51. Overexpression of seleno221subtilisin E and seleno63-subtilisin E in the presence of selenocysteine was performed as already described for (Se)2 –thioredoxin. The proteins, which were produced as inclusion bodies, were also purified on a CM-Sephadex-50 cation exchange column and renatured by dialyzing stepwise as described previously. 2.4. Expression and purification of wild-type subtilisin E and two thiolsubtilisin E Plasmid I (S221C), plasmid II (S63C), and plasmid III (wild type) were transferred to E. coli strain BL21DE3. The three proteins thiol221-subtilisin E, thiol63-subtilisin E, and wild-type subtilisin E were expressed and purified according to the method of Li and Inouye. 2.5. Refolding of proteins Purified proteins were dissolved in PBS buffer and dialyzed stepwise against a refolding buffer (50 mM Tris-HCl, pH 7.0, 1 mM CaCl2, 0.5 M (NH4)2SO4) containing decreasing amounts of urea (4, 2, 1, 0.5, and 0 M). After about 1 week incubation at 4 °C, the final refolding protein with no urea was applied to a Sephadex G-100 column to separate it from the unfolding protein.

T. Wang et al. / Journal of Catalysis 359 (2018) 27–35

2.6. Electrospray mass spectrometry analysis All MALDI-TOF mass spectra were acquired on a Voyager DESTR biospectrometry workstation (PerSeptive Biosystems, Framingham, MA) using a nitrogen laser (337 nm). The protein samples were prepared using a conventional dried droplet protocol in which sinapinic acid was used as the matrix. The matrix was prepared as a saturated aqueous solution that contained 70% acetonitrile and 0.3% trifluoroacetic acid. An aliquot sample of 1 ll was mixed with 10 ll of sinapinic acid matrix before 1.2 ll of the sample matrix mixture was deposited on the MALDI sample stage.

temperatures from 15 to 45 °C to determine the optimal temperature for the selenosubtilisin E-catalyzed reduction of hydroperoxide. 2.12. Kinetics of seleno63-subtilisin E The kinetic experiments were accomplished in MES buffer (pH 5.5) by changing one substrate’s concentration while keeping the other constant. Saturation kinetics was observed for the peroxidase reaction at all individual concentrations of ArSH and H2O2.

2.7. Circular dichroism measurements

3. Results and discussion

Circular dichroism (CD) spectra of the refolding proteins for wild-type subtilisin E, seleno221-subtilisin E, and seleno63subtilisin E were monitored by a JASCO J-810 circular dichroism system from 200 to 260 nm. Protein concentration was 10 lM, and a path length of 0.1 cm was used. For CD studies, two selenosubtilisin E in 50 mM pH 7.0, containing 0.5 M (NH4)2SO4 and 1 mM CaCl2, were used. Solutions were filtered through a 0.22 lm filter before measurement.

3.1. Rational redesign of the active site

2.8. Binding experiments Thiol63-subtilisin E and thiol221-subtilisin E were applied to the binding experiments. Determination of thiol substrate ArSH (3-car boxy-4-nitroben-zenethethiol) binding of the enzyme was carried out according to the method of Hildebrand and Benesi. 2.9. Assay of glutathione peroxidase activity The activities of enzymes were measured by a Shimadzu UV2450 UV/visible spectrophotometer. The GPx activities of enzymes were evaluated in two different reaction systems. One used ArSH as the enzyme’s substrate and 100 mM 4morpholineethanesulfonic acid (MES) containing 1 mM CaCl2 and 1 mM EDTA as the buffer (pH 5.5). The initial rates for the reduction of ROOH by ArSH were determined at 25 °C by monitoring the disappearance of ArSH at 410 nm. The other system was according to Wilson’s method. The reaction was carried out at 37 °C in 1 ml of a solution consisting of 50 mM Tris-HCl buffer (pH 7.0, containing 1 mM CaCl2 and 0.5 M (NH4)2SO4), 1 mM GSH, 1 unit of GSH reductase, and 1 lM enzyme. The mixture was preincubated for 5 min, and 0.25 mM NADPH solution was added. After the mixture was incubated for 3 min at 37 °C, the reaction was initiated by addition of 0.5 mM H2O2. The activity was determined by the decrease of NADPH absorption at 340 nm. 2.10. Assay of hydrolase activity The hydrolase activities of enzymes were estimated using 4nitrophenyl acetate as the enzyme’s substrate and 50 mM TrisHCl containing 1 mM CaCl2 and 0.5 M (NH4)2SO4 as the buffer (pH 7.0). The initial rates for the reaction of enzymes and substrate were determined at 37 °C by monitoring the appearance of 4nitrophenyl acetate at 400 nm. 2.11. Determination of optimal pH and temperature for seleno63subtilisin E catalysis The initial rates were measured using 100 lM ArSH and 0.5 mM hydrogen peroxide. The pH value of the buffer was changed from 5 to 8 to determine the initial rates of the reaction to obtain the optimal pH condition for selenosubtilisin E-catalyzed reaction. Similarly, a catalytic reaction was carried out at different

29

One of the most significant challenges of rebuilding and optimizing enzyme properties is rational design and redesign starting from the active site [17]. Computational techniques have graduated to an indispensable subsidiary instrumentality. To broaden substrate specificities and improve trace activities of selenosubtilisin with structure prediction by computer-aided molecular simulation, we redesigned the active site of selenosubtilisin by transferring the catalytically essential Sec to the edge of the substrate-binding pocket of the enzyme, as shown in Fig. 1. This minimum energy state made the thiol group of the substrate ArSH turn outward to approach the essential catalytic site Sec63, which was located at the border of the binding pocket in selenosubtilisin. Focusing on the interaction of substrate ArSH and the correlative amino acid residues, when the catalytic essential Sec was changed from position 221 to 63, ArSH had a slight shift from the bottom to the outside. The orientation of the SH group of ArSH was changed from an inward to an outward direction of the active-site cleft and the position of ArSH was moved outward about 3 Å (from 9.23 to 6.22 Å) to the border of the pocket. 3.2. Preparation of seleno63-subtilisin E Based on the assumption that efficient charging of tRNACys occurred with selenocysteine when cysteine was omitted, Cys could be replaced efficiently with Sec when Cys was omitted in the growth medium in a cysteine auxotrophic strain. The active residue Ser-63 of subtilisin E was mutated to Cys and then converted to Sec during expression in the auxotrophic strain. The procedure of purification of seleno63-subtilisin E from inclusion bodies and refolding it in vitro was similar to that for seleno221-subtilisin E. The subtilisin E was composed of an 8 kDa propeptide and a 28 kDa mature protein. Judged by SDS-PAGE (Fig. 2), the molecular mass of the purified protein before renaturation was estimated as 36 kDa (lane 4). After refolding in vitro, the propeptide of seleno63-subtilisin E was cleaved and degraded. The molecular mass of the renatured seleno63-subtilisin E was 28 kDa (lane 5). This result was the same as that for wild-type subtilisin E (lanes 2 and 3). In contrast, the mass of seleno221-subtilisin E was kept at 36 kDa before and after refolding for the propeptide of seleno63-subtilisin E that had not been cleaved (lanes 6 and 7). The different renaturation results for these two selenosubtilisins were due to the fact that the novel mutant (Sec63) protein retained a hydrolase capability that could result in the cleavage and degradation of the propeptide, whereas the mutant (Sec221) protein lost the hydrolase activity. Polyacrylamidegel (15%) was stained with Coomassie Brilliant Blue R-250. Lane 1 represents size standards. Each molecular mass is shown by pointing to the band: three purified proteins of wildtype subtilisin E (lane 2), seleno63-subtilisin E (lane 4), and seleno221-subtilisin E (lane 6) before refolding in vitro were the

30

T. Wang et al. / Journal of Catalysis 359 (2018) 27–35

Fig. 1. Final conformation of the substrate ArSH binding to seleno221-subtilisin E (A) and seleno63-subtilisin E (B) with Insight II/binding-site module (after removal of the residue His64). Green is the residue 63 at the edge of the substrate-binding pocket of the enzyme and turquoise is the residue 221 at the bottom.

seleno63-subtilisin E was cleaved and degraded in renaturation in vitro but that of the seleno221-subtilisin E was not. According to the modified Hildebrand–Benesi equation, the binding constant of the novel enzyme molecule thiol63-subtilisin E with substrate molecule ArSH was 1500 M
1 at pH 5.5, which was a slight increase from that of thiol221-subtilisin E, 1339 M
1. It is suggested that this shifting out of the essential catalytic site from the bottom of the binding pocket cannot markedly improve the affinity of substrate ArSH binding to the enzyme. 3.4. Peroxidase activity of seleno63-subtilisin E

Fig. 2. SDS-PAGE of purified and renatured seleno63-subtilisin E.

propeptide-subtilisin E complexes (36KD); after refolding by stepwise dialysis, both wild-type subtilisin E and seleno63-subtilisin E were both renatured into the mature proteins (28KD) without the propeptide (lane 3 and lane 5); conversely, the molecular mass of seleno221-subtilisin E was kept at 36 kDa after refolding (lane 7) because the propeptide had not been cleaved off.

3.3. Characterization of the selenoenzymes The purified and refolded selenoenzyme samples were characterized by MALDI-TOF mass spectrometry and circular dichroism spectroscopy. The MALDI-TOF mass spectrometry results (Fig. 3) revealed that the mass of seleno63-subtilisin E was 27,826 Da; the protein seleno221-subtilisin E as a control was 36,196 Da. This result was consistent with the molecular mass judged by SDSPAGE (Fig. 2), and both of the two selenoenzymes were predicted for the ESeO2H form. The secondary structure of seleno63-subtilisin E was analyzed by CD spectroscopy (Fig. 4). The spectrum of the seleno63subtilisin E (filled circles) was very similar to that of the wildtype subtilisin E (open circles). Only a very slight increase in the negative ellipticity at 219 nm was displayed for the seleno63subtilisin E. This suggested that the Ser63Sec substitution did not dramatically alter the secondary structure of the protein domain. The spectra were also very similar to those of the seleno221subtilisin E (filled squares). The parallel decrease in negative ellipticity from 210 to 220 nm suggested that the propeptide of the

The peroxidase activity of seleno63-subtilisin E and seleno221subtilisin E and some other catalysts is listed in Table 2. The GPx activities of the wild-type subtilisin E and the two seleniumcontaining enzymes were estimated under the same conditions [18]. The wild-type subtilisin E used as a control exhibited no GPx activity, whereas the purified renatured seleno63-subtilisin E displayed remarkably enhanced activity, at least 60-fold to 180fold that of seleno221-subtilisin E. Compared with diphenyl diselenide (PhSeSePh) under the same conditions, the GPx activity of this engineered enzyme was strikingly augmented from 21,000fold to 370,000-fold by shifting Sec from position 221 to 63. From the values of the GPx activities of these two seleniumcontaining enzymes, the initial velocity (v0) for 1 lM seleno221subtilisin E-catalyzed reduction of 0.25 mM cumene hydroperoxide (CuOOH) by 100 lM ArSH (20.4 ± 0.1 lmol min
1 lmol
1) was fivefold that of 0.25 mM H2O2 (4.0 ± 0.2 lmol min
1 lmol
1) and twofold that of 0.25 mM t-BuOOH (10.4 ± 0.1 lmol min
1 lmol
1). It was possible that CuOOH was able to take some binding advantage of the subtilisin protein template inside its hydrophobic pocket. It was interesting that this advantage of CuOOH binding to seleno63-subtilisin E seemed not to be significantly different from that to seleno221-subtilisin E when the essential catalytic position was changed to the border of the binding pocket. v0 for seleno63subtilisin E-catalyzed reduction of CuOOH by ArSH (1164 ± 2 lmol min
1 lmol
1) was 1.6-fold that of reduction of H2O2 (712 ± 4 lmol min
1 lmol
1) and 1.4-fold that of t-BuOOH (827 ± 1 lmol min
1 lmol
1). This is believed to reflect the fact that in the reengineered enzyme seleno63-subtilisin E the selenolate sits at the border of the binding pocket and may react with any approaching hydroperoxide: the enzyme has no real substrate specificity for hydroperoxides, provided that steric hindrance does not prevent their reaction.

31

T. Wang et al. / Journal of Catalysis 359 (2018) 27–35

Fig. 3. MALDI-TOF electrospray mass spectrometry analysis of seleno63-subtilisin E and seleno221-subtilisin E.

Table 2 Values of GPx activities of seleno63-subtilisin E compared with seleno221-subtilisin E and other catalysts. Catalysts

Thiol (GSH/ ArSH)

Hydroperoxide (ROOH)

Activity (lmol min
1 lmol
1)a

Ebselenb

GSH

H2O2 t-BuOOH CuOOH

0.99 0.35 1.28

PhSeSePhc

GSH ArSH

H2O2 H2O2

1.95 0.0012

Wild-subtilisin E

GSH/ArSH

H2O2

NDd

Semisynthetic selenosubtilisinee

ArSH

H2O2

2.1

GSH

H2O2

NDd

ArSH

H2O2 t-BuOOH CuOOH H2O2

4.0 10.4 20.4 NDd 712.0 827.0 1164.0 47.2 1445.0

Seleno

221

-subtilisin E

GSH Fig. 4. Far-UV CD spectra of renatured seleno63-subtilisin E. Seleno63-subtilisin E purified by CM-Sephadex C-50 column chromatography was renatured by stepwise dialysis as described under Section 2. Open circles, CD spectrum of wild-type subtilisin E; filled circles, that of the seleno63-subtilisin E; filled squares, that of the seleno221-subtilisin E with propeptide complex.

It was exciting that, besides the enhanced GPx activity used ArSH as the substrate, the GPx activity of enzyme-catalyzed reduction of hydroperoxides by GSH emerged [22]. At 37 °C and pH 7.0, v0 for reduction of 1 mM H2O2 by 0.5 mM GSH in the presence of 1 lM seleno63-subtilisin E was (47.2 ± 0.2)  10
6 M min
1, whereas the GPx activity for reduction of H2O2 by GSH with seleno221subtilisin E under the same conditions cannot be detected. Moreover, because the original catalytic triad of hydrolase (Asp32, His64, and Ser221) was preserved, the hydrolase activity of seleno63-subtilisin E remained equal to that of the wild-type subtilisin E and rather higher than that of seleno221-subtilisin E. The hydrolase activity values of wild-type subtilisin E, seleno63subtilisin E, and seleno221-subtilisin E were respectively 56.7 ± 0.5, 44 ± 1, and 2.5 ± 0.3 lmol min
1 mg
1. 3.5. The optimal pH and temperature for seleno63-subtilisin Ecatalyzed reduction of hydrogen peroxide by ArSH The GPx activity of seleno63-subtilisin E was examined over the pH range from 5 to 8 [23,24] and the temperature range from 15 to 45 °C. As shown in Fig. 5, pH and temperature profiles of seleno63subtilisin E activity were closer to those exhibited by natural GPx

Seleno63-subtilisin E

ArSH

GSH

H2O2 t-BuOOH CuOOH H2O2

Native (GPx, rabbit liver)f

GSH

H2O2

a The activities were determined as initial rates and corrected for the spontaneous reaction. All values were the means of at least five determinations, and the standard deviations were shown in parentheses. The conditions of the reactions are described in Section 2. b Values of GPx activity were reported previously [19]. c The GPx activity to GSH was reported in Wilson et al. and the GPx activity to ArSH was reported in Dong et al. [20]. d ND, no detectable GPx activity. e The catalyst was prepared as described in Wu and Hilvert [9], and the activity was detected in this study. f The GPx activity of the native GPx was reported in Mannervik [21].

[25] than those of seleno221-subtilisin E. Seleno63-subtilisin E had an optimal pH of 7.0, which was close to that of natural GPx (8.8), and a temperature optimum of 37 °C, which was lower than that of natural GPx (50 °C) [26]. The optimal pH and temperature for seleno221-subtilisin E were 5.5 and 25 °C, respectively. 3.6. steady-state kinetics and catalytic mechanism of seleno63subtilisin E As shown in Table 3, seleno221-subtilisin E catalyzed decomposition of t-BuOOH using ArSH as the substrate with a low GPx

32

T. Wang et al. / Journal of Catalysis 359 (2018) 27–35

Fig. 5. Effects of the catalytic conditions on the GPx activity for the reduction of H2O2 by ArSH catalyzed by seleno63-subtilisin E. (A) Initial velocity (v0) for the reaction versus pH. (B) Initial velocity (v0) for the reaction versus temperature. The activity was determined when the concentrations of ArSH and H2O2 were 100 and 500 lM, respectively.

Table 3 Comparison of the kinetic parameters for the GPx activity of the two selenosubtilisin E.a Selenosubtilisin E 221

Seleno -subtilisin E Seleno63-subtilisin E Ratio

kmax (min
1) 1675 3304 2

Kt-BuOOH (mM) 195 4 1/48

KArSH (lM)

kmax/Kt-BuOOH (M
1 min
1) 3

1.97  107 4.29  107 2

8.59  10 8.26  105 96

85 78 1

kmax/KArSH (M
1 min
1)

a Reactions were carried out in MES buffer (100 mM MES, 1 mM CaCl2, 1 mM EDTA, pH 5.5) at 25 °C. The experimental data were fitted to a ping-pong mechanism to obtain the parameters shown.

activity of 10.4 ± 0.1 lmol min
1 lmol
1 relative to the native GPx. After the catalytic essential group Sec shifted from the bottom to the edge of the binding pocket, the GPx activity of the resulting novel seleno63-subtilisin E was increased by almost two orders of magnitude in comparison with that of seleno221-subtilisin E. Moreover, this novel engineered enzyme accepted not only ArSH as a substrate but also GSH, which was the general substrate for the naturally occurring GPx [27]. This indicates that it is one of the important factors for improving the properties of existing enzymes by altering the catalytic essential active site of the enzyme for easier contact with the corresponding substrate, consequently enhancing the catalytic efficiency of the enzyme [28]. The initial velocity (v0) for t-BuOOH reduction of ArSH catalyzed by seleno63-subtilisin E was determined as a function of substrate concentration at 25.0 °C and pH 5.5 when the concentration of one substrate was varied and that of the other was fixed. Michaelis– Menten kinetics was observed under all conditions investigated [29]. Double reciprocal plots of the initial velocity versus substrate concentration showed the characteristic parallel lines (Fig. 6A and B) of a ping-pong mechanism, in analogy with those of seleno221-subtilisin E.

Table 4 Kinetic parameters for the GPx activity of seleno63-subtilisin E.a [ArSH] (lM)

kcat (app) (min
1)

Kt-BuOOH (app) (mM)

kcat (app)/Kt-BuOOH (app) (M
1 min
1)

25 50 100 200

1044 ± 62 1503 ± 80 2245 ± 99 2986 ± 113

1.26 ± 0.05 1.82 ± 0.1 2.72 ± 0.2 3.61 ± 0.2

(8.26 ± 0.05)  105 (8.26 ± 0.02)  105 (8.25 ± 0.01)  105 (8.26 ± 0.01)  105

[t-BuOOH] (M)

kcat (app) (min
1)

KArSH (app) (lM)

kcat (app)/KArSH (app) (M
1 min
1)

0.05 0.1 0.2 0.4

1445 ± 87 2052 ± 94 2730 ± 120 3304 ± 13

33.76 ± 0.5 48.06 ± 0.8 63.64 ± 1.2 77.02 ± 1.2

(4.28 ± 0.04)  107 (4.27 ± 0.02)  107 (4.29 ± 0.01)  107 (4.29 ± 0.01)  107

a The reactions were carried out in MES buffer (100 mM MES, 1 mM CaCl2, 1 mM EDTA, pH 5.5) at 25 °C.

The kinetic parameters for the reactions between ArSH and tBuOOH catalyzed by seleno63-subtilisin E are detailed in Table 4. These values were deduced by fitting the experimental data to

Fig. 6. Double-reciprocal plots for the reduction of t-BuOOH by ArSH catalyzed by seleno63-subtilisin E. (A) [E]0/v0 versus 1/[ArSH] (mM
1) at [t-BuOOH] = 0.05 M (j), 0.1 M (d), 0.2 M (N), and 0.4 M (.). (B) [E]0/v0 versus 1/[t-BuOOH] (M
1) at [ArSH] = 25 lM (j), 50 lM (d), 100 lM (N), and 200 lM (.).

T. Wang et al. / Journal of Catalysis 359 (2018) 27–35

33

Fig. 7. Plots of the initial velocity V0 (lM min
1) vs. [ArSH] and [t-BuOOH] for the enzymatic peroxidase reaction of seleno63-subtilisin E (blue line) compared with that of seleno221-subtilisin E (red line). The data were fitted to a saturation curve. (A) The initial velocity V0 vs. [ArSH] of seleno221-subtilisin E were enough lower than that of seleno63-subtilisin E so that the saturation curve is an approximate line that is amplified in the inset. (B) Similarly, the initial velocity V0 vs [t-BuOOH] of seleno63-subtilisin E were enough higher than that of seleno221-subtilisin E so that the saturation curve is an approximate line that is reduced in the small frame.

Scheme 1. The interactions between the prosthetic group and the residues in subtilisin’s oxyanion hole.

a ping-pong kinetic scheme. The relevant steady-state rate equation is

V0 kmax ½ArSH½t-BuOOH ¼ ½E K ArSH ½t-BuOOH þ K t-BuOOH ½ArSH þ ½t-BuOOH½ArSH where kmax is a pseudo-first-order rate constant and Kt-BuOOH and KArSH are the Michaelis constants for the peroxide and thiol, respectively. The first-order rate constant kcat (app) and the apparent Michaelis constant Km t-BuOOH at 100 lM ArSH were determined to be 2245 min
1 and 2.72 mM, respectively. At 100 lM tBuOOH, KmArSH was 4.8  10
5 M. The apparent second-order rate constants kcat/Km t-BuOOH and kcat/KmArSH provided measures of the rates of reactions between the free enzyme and the relative substrates (ArSH and hydrogen peroxide, respectively). kcat/KmArSH of seleno63-subtilisin E was 4.3  107 M
1 min
1 and kcat/Kmt-BuOOH was 8.3  105 M
1 min
1. kcat/KmArSH of seleno63-subtilisin E was of the same order of magnitude as that of natural GPx, indicating that they had similar selectivity to ArSH. Although kcat/Kmt-BuOOH of seleno63-subtilisin E was still three orders of magnitude lower than that of natural GPx (108 M
1 min
1), it was much higher than those of most GPx mimics (for example, 4.5  103 M
1 min
1 for the selenium-containing catalytic antibody Se-4A4) [30]. In a comparison of the kinetic parameters for the GPx activity [31] of seleno63-subtilisin E with the parameters for that of seleno221-subtilisin E, the second-order rate constant kmax/Kt-BuOOH increased 96-fold and kmax/KArSH increased twofold in Table 3. The question arose of why the alteration of the catalytic site from the bottom position 221 to the 63 at the edge of the binding pocket caused a massive increase in catalytic efficiency. We considered one of the reasons to be that by this reengineering of the selenosubtilisin active site, some steric hindrance for the attack

of the other molecule substrate t-BuOOH upon the selenolate was relieved. In seleno221-subtilisin E, the selenium side chain was buried in a deep pocket and was presumably less accessible to hydroperoxides than its natural counterpart. After the essential catalytic Sec was changed from the bottom position 221 to the 63 at the edge of the binding pocket, the mode of the substrate ArSH binding to the enzyme changed with it. As shown in Fig. 1, the orientation of the SH group of ArSH was changed from the inward to the direction of the active site cleft and the location of ArSH was moved outward about 3 Å (from 9.23 to 6.22 Å) to the border of the pocket. This behavior would give it enough room for the substrates to enter and the products to leave in the active site, consequently improving the catalytic velocity. In addition, selenosubtilisin is a peroxide-dependent enzyme. We can discover in Fig. 7A and B that the initial velocities of seleno63-subtilisin E-catalyzed reduction of t-BuOOH by ArSH were enormously more rapid than those for seleno221-subtilisin E. According to the steady-state rate equation [32], if the maximal kcat value for the enzymic reaction (kmax) is altered, then the peroxide substrate must be involved in a step that is at least partially rate-determining. In seleno221-subtilisin E, considered in Scheme 1, the strong interactions between the prosthetic group and the residues in subtilisin’s oxyanion hole and His64 may stabilize the selenolate anion firmly. The conversion of ESeSAr to ESe- may be expected to be the rate-limiting step because the selenolate anion was so firm that it would hamper the release of the reaction product, resulting in lost catalytic efficiency. After the Sec was altered from the bottom position 221 to 63, and the His64 side chain was protonated, the rate-determining step may be expected to be conversion of (ESe- + R’OOH) to (ESeOH + R’OH) because the conversion of ESeSAr to ESe- could be far more rapid, as the selenolate residue was slightly destabilized. Thus, the catalytic activity of selenosubtilisin should be enhanced to mend the velocity for the transformation of ESe- to EseOH. Fortunately, this selective mutant on the border of the binding pocket provided more opportunities to relieve some steric hindrance of hydroperoxide t-BuOOH attack upon the selenolate and slightly destabilize the selenolate residue. Indeed, the redesign of the genetically engineered enzyme offers mechanistic insights that may facilitate the rational improvement of the enzyme’s efficiency in this study [33]. The balance between stabilizing a high-energy species and lowering its energy to the point of unreactivity has been perfected. This also provides a strategy for designing improved substrates of the enzyme and optimizing its catalytic efficiency through additional site-specific changes of catalytic factors within the active site. This success can be anticipated to open the door to changing

34

T. Wang et al. / Journal of Catalysis 359 (2018) 27–35

the active site by changing the catalytic selenium position in further structure–function research on biologically important selenoenzymes. 4. Conclusions Previous attempts at generating and engineering selenoenzymes with peroxidase activity always used chemical or genetic incorporation of a prosthetic group into an existing protein binding site. However, these methods do not achieve a perfect outcome, presumably because the active sites of intrinsic preexisting protein scaffolds are in favor of enzyme binding of the corresponding counterpart when the novel enzyme activities emerge simultaneously. In this study, we have taken advantage of specifically sitedirected targeting of Sec into proteins using a cysteine auxotrophic expression system, combined with protein structure prediction via computational molecular simulation, to rationally redesign the active site of subtilisin to gain remarkably enhanced GPx catalytic efficiency. In view of the superiority of the genetic engineering method, we try to find a suitable site in the active site of the enzyme to facilitate the enzyme reacting with the corresponding substrate molecules and improve the catalytic efficiency of the enzyme. Combined with computer simulation of protein molecular structure, based on the orientation of enzymes and substrate molecules and the energy minimum principle, we found an optimal position at the pocket edge of subtilisin. The success in shifting the essential catalytic group to a more suitable position provided an approach that could be applied to the generation and rebuilding of other efficient selenium-containing selenoenzymes for more perfect properties. It also gave us the opportunity to establish insight into pursuing detailed study to search for a better enzyme model by chemically or genetically modifying naturally occurring proteins, and to understand the mechanism of enzymes and enzymelike catalysts. Acknowledgments We acknowledge financial support from the National Natural Science Foundation of China (Nos. 21420102007, 21574056, and 91527302) and the Chang Jiang Scholars Program of China. The authors declare no competing financial interest. References [1] (a) O. Epp, R. Ladenstein, A. Wendel, The refined structure of the selenoenzyme glutathione peroxidase at 0.2-nm resolution, Eur. J. Biochem. 133 (1983) 51–69; (b) X.M. Liu, L.A. Silks, C.P. Liu, M. Ollivault-Shiflett, X. Huang, J. Li, G.M. Luo, Y. M. Hou, J.Q. Liu, J.C. Shen, Incorporation of tellurocysteine into glutathione transferase generates high glutathione peroxidase efficiency, Angew. Chem. Int. Ed. 48 (2009) 2020–2023. [2] (a) P. Saura, R. Suardíaz, L. Masgrau, J.M. Lluch, À. Gonzalez-Lafont, Unraveling how enzymes can use bulky residues to drive site-selective C-H activation: the case of mammalian lipoxygenases catalyzing arachidonic acid oxidation, ACS Catal. 4 (2014) 4351–4363; (b) J.B. Siegel, A. Zanghellini, H.M. Lovick, G. Kiss, A.R. Lambert, J.L. Gallaher, D. Hilvert, M.H. Gelb, B.L. Stoddard, K.N. Houk, F.E. Michael, D. Baker, Computational design of an enzyme catalyst for a stereoselective bimolecular diels-alder reaction, Science 329 (2010) 309–313. [3] G. Ma, W. Zhu, H. Su, N. Cheng, Y. Liu, Uncoupled epimerization and desaturation by carbapenem synthase: mechanistic insights from QM/MM studies, ACS Catal. 5 (2015) 5556–5566. [4] (a) M.D. Toscano, K.J. Woycechowsky, D. Hilvert, Minimalist active-site redesign: teaching old enzymes new tricks, Angew. Chem. Int. Ed. 46 (2007) 3212–3236; (b) G. Kiss, N. Çelebi-Olcum, R. Moretti, D. Baker, K.N. Houk, Computational enzyme design, Angew. Chem. Int. Ed. 52 (2013) 5700–5725; (c) L. Jiang, E.A. Althoff, F.R. Clemente, L. Doyle, D. Rothlisberger, A. Zanghellini, J.L. Gallaher, J.L. Betker, F. Tanaka, C.F. Barbas, D. Hilvert, K.N. Houk, B.L. Stoddard, D. Baker, De novo computational design of retro-aldol enzymes, Science 319 (2008) 1387–1391.

[5] (a) C.X. Hou, Q. Luo, J.L. Liu, L. Miao, C.Q. Zhang, Y.Z. Gao, X.Y. Zhang, J.Y. Xu, Z. Y. Dong, J.Q. Liu, Construction of GPx active centers on natural protein nanodisk/nanotube: a new way to develop artificial nanoenzyme, ACS Nano 6 (2012) 8692–8701; (b) G. Mugesh, H.B. Singh, Synthetic organoselenium compounds as antioxidants: glutathione peroxidase activity, Chem. Soc. Rev. 29 (2000) 347–357; (c) K.P. Bhabak, G. Mugesh, Functional mimics of glutathione peroxidase: bioinspired synthetic antioxidants, Acc. Chem. Res. 43 (2010) 1408–1419. [6] (a) B.B. Lin, G.C. Ma, Y.J. Liu, Mechanism of the glutathione persulfide oxidation process catalyzed by ethylmalonic encephalopathy protein 1, ACS Catal. 10 (2016) 7010–7020; (b) J.Y. Rotruck, A.L. Pope, H.E. Ganther, A.B. Swanson, D.G. Hafeman, W.G. Hoekstra, Selenium: biochemical role as a component of glutathione peroxidase, Science 179 (1973) 588–590; (c) H.E. Ganther, R.J. Kraus, Oxidation states of glutathione peroxidase, Methods Enzymol. 107 (1984) 593–602. [7] (a) X. Huang, Z.Y. Dong, J.Q. Liu, S.Z. Mao, J.Y. Xu, G.M. Luo, J.C. Shen, Seleniummediated micellar catalyst: an efficient enzyme model for glutathione peroxidase-like catalysis, Langmuir 3 (2007) 1518–1522; (b) F.F. Chu, J.H. Doroshow, R.S. Esworthy, Expression, characterization, and tissue distribution of a new cellular selenium-dependent glutathione peroxidase, GSHPx-GI, J. Biol. Chem. 268 (1993) 2571–2576. [8] A.A. Vernekar, G. Mugesh, Catalytic reduction of graphene oxide nanosheets by glutathione peroxidase mimetics reveals a new structural motif in graphene oxide, Chem. Eur. J. 49 (2013) 16699–16706. [9] Z.P. Wu, D. Hilvert, Selenosubtilisin as a glutathione peroxidase mimic, J. Am. Chem. Soc. 112 (1990) 5647–5648. [10] R. Syed, Z.P. Wu, J.M. Hogle, D. Hilvert, Crystal structure of selenosubtilisin at 2.0-A resolution, Biochemistry 32 (1993) 6157–6164. [11] K.L. House, R.B. Dunlap, J.D. Odom, Z.P. Wu, D. Hilvert, Structural characterization of selenosubtilisin by selenium-77 NMR spectroscopy, J. Am. Chem. Soc. 112 (1992) 8573–8579. [12] J. Li, X.M. Liu, Y.T. Ji, Z.H. Qi, Y. Ge, J.Y. Xu, J.Q. Liu, G.M. Luo, J.C. Shen, Biosynthesis of selenosubtilisin: a novel way to target selenium into the active site of subtilisin, Chin. Sci. Bull.. 16 (2008) 2454–2461. [13] G. Takebe, J. Yarimizu, Y. Saito, T. Hayashi, H. Nakamura, J. Yodoi, S. Nagasawa, K. Takahashi, Glutathione peroxidase family and selenoprotein P, J. Biol. Chem.. 277 (2002) 41254–41258. [14] P. Liu, J.A. Rodriguez, Y. Takahashi, K. Nakamura, Water-gas-shift reaction on a Ni2P (001) catalyst: Formation of oxy-phosphides and highly active reaction sites, J. Catal. 262 (2009) 294–303. [15] G.M. Luo, Z.Q. Zhu, L. Ding, S.J. Gao, Q.A. Sun, Z. Liu, T.S. Yang, J.C. Shen, Generation of selenium-containing abzyme by using chemical mutation, Biochem. Biophys. Res. Commun. 198 (1994) 1240–1247. [16] H.A. Benesi, J.H. Hildebrand, A spectrophotometric investigation of the interaction of iodine with aromatic hydrocarbons, J. Am. Chem. Soc. 71 (1949) 2703–2707. [17] I. Axarli, A.W. Muleta, E.G. Chronopoulou, A.C. Papageorgiou, N.E. Labrou, Directed evolution of glutathione transferases towards a selective glutathionebinding site and improved oxidative stability, BBA Gen. Subj. 1 (2017) 3416– 3428. [18] F. Ursini, M. Maiorino, R. Brigelius-Flohe, K.D. Aumann, A. Roveri, D. Schomburg, L. Flohe, Diversity of glutathione peroxidases, Methods Enzymol. 252 (1995) 38–53. [19] S.R. Wilson, P.A. Zucker, R.R.C. Huang, A. Spector, Development of synthetic compounds with glutathione peroxidase activity, J. Am. Chem. Soc. 111 (1989) 5936–5939. [20] Z.Y. Dong, J.Q. Liu, S.Z. Mao, X. Huang, B. Yang, X.J. Ren, G.M. Luo, J.C. Shen, Aryl thiol substrate 3-carboxy-4-nitrobenzenethiol strongly stimulating thiol peroxidase activity of glutathione peroxidase mimic 2, 2’-ditellurobis(2deoxy-b-cyclodextrin), J. Am. Chem. Soc. 50 (2004) 16395–16404. [21] B. Mannervik, Glutathione peroxidase, Methods Enzymol. 113 (1985) 490– 495. [22] W. Wei, C. Wang, Y. Zhao, S.C. Peng, H.Y. Zhang, Y.P. Bian, H.X. Li, X.G. Zhou, H. Li, Biochemical composite synthesized by stepwise crosslinking: An efficient platform for one-pot biomass conversion, J. Catal. 327 (2015) 78–85. [23] J. Zhang, J. Li, J. Zhang, R. Xie, W. Yang, Aqueous synthesis of ZnSe nanocrystals by using glutathione as ligand: the pH-mediated coordination of Zn2+ with glutathione, J. Phys. Chem. C 114 (2010) 11087–11091. [24] A. Krezel, W. Bal, Studies of zinc (II) and nickel (II) complexes of GSH, GSSG and their analogs shed more light on their biological relevance, Bioinorg. Chem. Appl. 2 (2004) 293–305. [25] Y. Yamamoto, K. Takahashi, Glutathione peroxidase isolated from plasma reduces phospholipid hydroperoxides, Arch. Biochem. Biophys. 305 (1993) 541–545. [26] L. Flohe, B. Eisele, A. Wendel, Glutathione peroxidase. I. isolation and determinations of molecular weight, Hoppe-Seyler’s, Z. Physiol. Chem. 352 (1971) 151–158. [27] L. Liu, S.Z. Mao, X.M. Liu, X. Huang, J.Y. Xu, J.Q. Liu, G.M. Luo, J.C. Shen, Functional mimicry of the active site of glutathione peroxidase by glutathione imprinted selenium-containing protein, Biomacromolecules 1 (2008) 363– 368. [28] I.M. Bell, D. Hilvert, Peroxide dependence of the semisynthetic enzyme selenosubtilisin, Biochemistry 32 (1993) 13969–13973.

T. Wang et al. / Journal of Catalysis 359 (2018) 27–35 [29] I.M. Bell, M.L. Fisher, Z.P.D. Wu, Hilvert, Kinetic studies on the peroxidase activity of selenosubtilisin, Biochemistry 32 (1993) 3754–3762. [30] L. Ding, Z. Liu, Z.Q. Zhu, G.M. Luo, D.Q. Zhao, J.Z. Ni, Biochemical characterization of selenium-containing catalytic antibody as a cytosolic glutathione peroxidase mimic, Biochem. J. 332 (1998) 251–255. [31] R. Gounder, M.E. Davis, Monosaccharide and disaccharide isomerization over Lewis acid sites in hydrophobic and hydrophilic molecular sieves, J. Catal. 308 (2013) 176–188.

35

[32] S. Toppo, L. Flohe, F. Ursini, S. Vanin, M. Maiorino, Catalytic mechanisms and specificities of glutathione peroxidases: variations of a basic scheme, BBA Gen. Subj. 11 (2009) 1486–1500. [33] H. Renata, R.D. Lewis, M.J. Sweredoski, A. Moradian, S. Hess, Z.J. Wang, F.H. Arnold, Identification of mechanism-based inactivation in p450-catalyzed cyclopropanation facilitates engineering of improved enzymes, J. Am. Chem. Soc. 38 (2016) 12527–12533.